Epilepsy Essentials: Practical Uses of EEG in the Operating Room

Clinical neurophysiologic procedures are important ancillary tools used during surgery involving the nervous system. Designed to improve postoperative outcomes and limit perioperative morbidity and mortality rates, clinical neurophysiologic procedures are applied during surgeries when they may involve or affect cerebral function. EEG is an adjunctive method used to obtain both baseline and dynamic electrophysiologic data during surgical procedures involving the brain and provides real-time information on functional changes, aiding in intraoperative monitoring and decision-making. Dynamic electrocorticography (ECoG) is increasingly used as a resource during high-risk surgeries. By observing electrophysiologic function, practitioners can recognize an impending direct or indirect injury pattern early and prevent irreversible damage involving the nervous system.¹

Electrocorticography
ECoG, which measures electrocerebral activity directly from the cerebral cortex and subcortex, is commonly used in operating rooms and epilepsy monitoring units.^1,2 In ECoG, electrode contacts are placed directly on the brain parenchyma to record EEG data. 

During baseline ECoG recording, the frequency and location of epileptiform discharges are important to assess, because surgery may change them. Frequent epileptiform activity could incrementally affect the static nature of a baseline neurologic impairment, which may complicate the separating of newly acquired deficits from existent baseline deficits present preoperatively. When epileptiform activity is widespread during simultaneous direct electrical stimulation (DES) and functional brain mapping (FBM), observed neurologic deficits could overestimate the location of eloquent cortex–involving critical functions because of involvement of a broader neuronal network that is activated outside the site of electrical stimulation.^3,4,5 This could result in a subtotal resection and worsen the prognosis, leading to an unfavorable long-term neurologic outcome.^3,6

ECoG has been used both diagnostically and prognostically to predict the outcome of epilepsy surgery.^5-7 However, despite the potential advantages of the provision of dynamic information and immediate feedback to the surgeon, reports have shown inconsistent benefits of using ECoG alone to guide resections except in patients with focal cortical dysplasia,² in whom uniquely abnormal patterns are present. Focal cortical dysplasia is frequently associated with bizarre morphologies on ECoG involving chaotic high-frequency epileptiform activity.5 Furthermore, resection of brain tissue demonstrating interictal epileptiform activity during epilepsy surgery based solely on ECoG readings may not render better outcomes, with resection of electrographic seizure foci being more beneficial in tailoring epilepsy surgery to target the epileptogenic zone.²

Intraoperative ECoG is time-limited to minimize exposure of the brain, and ictal recordings are typically not obtained reliably. When ictal recording is necessary to define the seizure onset zone, an extraoperative evaluation in a dedicated epilepsy monitoring unit with implantation of subdural and stereotactic intracranial EEG electrodes is required. Newer methods of recording ECoG are being developed.^3,6,7 This may improve its accuracy and achieve better outcomes beyond those obtained with commercially available sensors with standard temporal and spatial arrangements (Figures 1A, 1B, 2A, and 2B).^3,4,6,7

Figure 1. A 6-contact subdural strip electrode array on exposed brain during awake craniotomy positioned near the site of planned resection (A) and a customized 22-channel hollow circular grid during direct electrical stimulation of the cortex during functional brain mapping through the aperture (B).

An awake craniotomy is a clinical procedure performed during neurosurgery to provide dynamic information and test the integrity of the nervous system to identify emerging neurologic deficits linked to the area of the brain being operated on.⁴ Performing a craniotomy on an awake patient in the operating room is beneficial when surgery targets are resected in or near anatomic regions potentially involving eloquent brain functions.³ ECoG is best used in concert with DES for FBM to identify essential areas that should be avoided during surgery and minimize the likelihood of postoperative complications.⁵

Electrocortical activity monitoring with ECoG not only can detect epileptiform activity but also can identify the presence or absence of afterdischarges caused by DES during FBM.^1-5 Afterdischarges provoked by DES may result in transient bursts of epileptiform activity being triggered and recorded on ECoG.⁴ Their importance lies in identifying the presence of an afterdischarge, its duration, and its evolution. When afterdischarges persist, they portend the potential to evolve into electroclinical seizures, signifying the need for urgent intervention.^4,5 The ramifications of an electroclinical seizure during operation may prevent the ability to perform further FBM, disrupt the procedure, lead to a prolonged postictal state, delay postoperative recovery, and, rarely, lead to injury.^1,3-5 Abortive methods include applying cold saline, performing a second DES, and administering a benzodiazepine or antiseizure medication. When prolonged afterdischarges are identified and reproducible, repeating DES at higher current amplitudes should be avoided, as they could trigger afterdischarges that evolve into electroclinical seizures.^4,5

Figure 2. Spontaneous focal periodic epileptiform discharges (A), an afterdischarge present in a subdural strip array following direct electrical stimulation of the brain (B), and a focal afterdischarge present following direct electrical stimulation of the brain within the aperture of the hollow circular grid (C). Note the 2 channels obscured by artifact on the strip electrode in (B) (channel 1 used as reference, with 1–6 separated by 1-cm interelectrode distances) and the restricted field involving 3 electrodes in (A) recorded using the hollow circular grid (channel 22 used as a reference, with 0.5-cm interelectrode distances). Parameters of recording: Display speed, 30 mm/s; filter settings, 1–100 Hz; 60-Hz notch ON; sensitivity 50 μV/mm. 

DES in eloquent areas of the brain that contain critical functions such as motor, language, vision, and sensation are essential to identify and avoid. These functions, when identified during FBM, signify areas that would lead to a neurologic deficit if resected. ECoG has the potential to define, monitor, and predict seizures that could result in brain injury, and is most useful when used in conjunction with DES for FBM to guide surgical resection and minimize the risk of perioperative morbidity.

Guidelines for DES include maximum safe charge density ranging from 5 to 7 μC/cm²/phase for subdural electrodes. Typical paradigms for DES used for FBM with subdural disc electrodes include bipolar square wave pulses involving 50-Hz frequency, 200- to 500-μs pulse duration, electrical trains between 2 and 8 seconds, and current strengths ranging from 1 to 20 mA. Mapping using stereoelectroencephalography commonly uses parameters such as 1 and 50 Hz frequencies, pulse widths ranging from 100 to 500 μs, train duration from 2 to 8 seconds, and current intensity ranging from 0.5 to 10 mA with charge densities that vary depending on the frequency and pulse duration of the electrical stimulation chosen.⁵ Previous studies using ECoG alone have been unable to consistently demonstrate benefit, although they are confounded by using the same devices and techniques, limiting favorable outcomes. Unique electrode arrays involving customized designs may demonstrate greater yields from ECoG.^2,6-8 Newer innovative designs involving novel geometries and higher-density electrode arrays may improve detection of epileptiform activity. By virtue of increased recording sensitivity, the value of ECoG can be enhanced and translate to successful outcomes using ECoG during neurosurgery.^6,7

Subcortical stimulation is less standardized compared with direct electrical cortical stimulation.⁵ White matter injury is a predictor of postoperative neurologic morbidity, and subcortical stimulation is an important tool that is used in evaluating tracts that are either involved in the pathology or close to the surgical target beneath the cortical surface. Subcortical FBM helps prevent neurologic deficits by protecting essential white matter fibers. Subcortical mapping can be performed concomitantly or sequentially after cortical stimulation once a surgical corridor is made to access the subcortical space.⁵ Examples of white matter tracts that are often evaluated include the arcuate fasciculus (language), inferior frontal-occipital fasciculus (language and vision), and corticospinal tract (motor, with motor evoked potentials), in addition to other regions of potentially eloquent regions of the brain.^9,10 The white matter tract being tested will guide the stimulation parameters used and correlate with the neurologic examination.^1,5

Carotid Surgery
Intraoperative EEG monitoring has been performed during carotid endarterectomy (CEA) with and without the use of evoked potentials to signal ischemia.¹ Endarterectomy usually involves intraluminal surgical removal of atherosclerotic plaque but may involve other vascular pathology involving large-vessel supply of blood flow to the brain. Because CEA involves the common carotid or a portion of the internal carotid artery, it requires transient cross-clamping of the carotid artery during the surgical procedure.^11,12 When blood flow is compromised, ischemia may occur in the ipsilateral cerebral hemisphere, which, when prolonged, can result in an ischemic stroke. 

Because patients are sedated with general anesthesia, clinical symptoms heralding ischemia may be masked. Using continuous EEG monitoring provides ongoing objective evidence during surgery that may predict ischemia and subsequent neurologic injury.¹³ When intraoperative EEG monitoring during carotid surgery identifies a change in baseline electrocerebral activity imposed by restricting cerebral blood flow in real time, the surgeon can intervene in an effort to prevent evolving ischemia. 

Intraoperative EEG findings may include ipsilateral hemispheric theta and delta slowing, loss of fast (beta) activity, or attenuation of background voltage activity—usually within ~30 seconds after cross-clamping—which may indicate the onset of reversible ischemia. When changes on the EEG associated with ischemia are recognized during surgery, surgeons may consider use of a shunt to maintain cerebral blood flow around the cross-clamp or release the clamp from the carotid artery to avoid prolonged ischemia that could result in irreversible neurologic injury.^11,13

Previously, EEG monitoring during CEA was a common practice; however, most surgeons no longer use universal monitoring due to limited availability, costs, validated efficacy from shunting, and surgeon preference.¹ Instead, physicians may reserve EEG monitoring during CEA for patients with clinical predictors for cerebral ischemia during CEA, such as female sex, history of diabetes, contralateral carotid high-grade stenosis or occlusion, or symptomatic carotid stenosis. In these cases, or where surgery has a high risk of cerebral ischemia, EEG monitoring may be desirable during carotid surgery.¹²

Depth of Anesthesia
The depth of sedation during a surgical procedure can be monitored on the basis of electrocerebral activity. Standardized anesthesia protocols are intended to prevent undersedation or oversedation. Monitoring can help ensure the correct amount of anesthesia is administered during surgery. This is crucial to ensure patient safety and efficacy of anesthesia because if the level of anesthesia is inadequate, patient recall or awareness during the procedure may occur. In addition, when anesthesia is excessive, unfavorable hemodynamic changes may occur. EEG monitoring can assist in precise dosing of anesthesia¹⁴ to limit perioperative complications and optimize recovery time to improve long-term postoperative outcomes.¹⁵ 

A common practice is to use processed EEG involving bispectral index (BIS) monitoring though raw EEG monitoring.¹⁶ BIS monitoring involves quantification of the level of anesthesia and reflects the potential for patient awareness during surgical procedures. Rapidly processed EEG is readily available, is in widespread use, and can be used to titrate and taper anesthesia given its ease of application and interpretation by anesthesiologists.¹⁷ BIS monitoring uses sensors placed on the forehead to process EEG frequencies to calculate a value between 0 (no brain activity) and 100 (normal), with a BIS value of 40 to 60 representing a level of sedation adequate for surgery. BIS monitoring has potential as an objective measure to prevent intraoperative recall while avoiding excessive sedation, and to predict ischemia when medication management proves ineffectual.¹⁷

The Future of EEG Monitoring in the Operating Room
Machine learning and artificial intelligence using software algorithms are increasingly designed to accurately identify epileptiform activity.¹⁸ Computer-based models have the potential to allow for more efficient and accurate interpretation of complex and high volumes of EEG data acquired during surgery. Advances in machine learning are likely to have many implications for the use of EEG in the operating room, which is hoped to lead to improved outcomes following surgeries that directly or indirectly involve the brain.¹⁸ Enhancement of processing and quantitative analyses of EEG data may also prove vital in guiding surgical procedures and anesthetic management.¹⁷

Conclusions
Intraoperative monitoring involves different applications during which EEG is used as a tool to monitor brain function during a surgical procedure. EEG monitoring acquires electrophysiologic signals in real time and provides immediate feedback to the surgeon during the procedure. Intracranial EEG using ECoG is a useful adjunct that provides an objective measure to detect and quantify epileptiform activity during brain surgery. Surface EEG monitoring during carotid surgery can predict the potential for brain injury from ischemia. EEG with BIS monitoring can help anesthesiologists regulate the depth of sedation from anesthesia. Earlier recognition of intraoperative issues allows intervention by the surgeon to alter the surgical procedure and prevent an unfavorable neurologic outcome. The increasing use of algorithms involving machine learning and artificial intelligence has the potential to improve the efficiency and accuracy of intraoperative EEG monitoring.